Even if you accept only the strictest possible definition of materialism, which is called physicalism and which posits that everything consists of elementary particles, this can bring you a long way. The point is that matter obtains only its most primitive properties from its elemental composition, by far most of the properties it exhibits stem from the way it is structured. A chemical example of this is that all organic substances consist of carbon, oxygen, and hydrogen, but the different shapes of their molecules allows for a vast variety of materials.

In the physicalist interpretation, an idea is a collection of neuronal or electronic configurations which - by mechanisms not yet entirely understood - can be transcribed into e.g. binary code or the Latin alphabet. We know a configuration represents the idea we want it to, if we can feed thus-configured solvers a problem to test that it has “understood” the idea, and it will put out the expected solutions. Great numbers of increasingly complex, successful experiments with neural networks and deep learning should serve to illustrate that such an interpretation of an idea is viable. In particular, the electromagnetic spectrum is any configuration which we can feed into our brain - perhaps in the form of reading a textbook or looking at diagrams, or into a machine - in the form of code - which then enables either to distinguish between different wavelengths of light.